Potting electronics...

whoizrob

I may be over thinking this but has anyone fully potted a handset?

I am thinking about possible soultions for DIY PP02 metering on my Rebreather design but most DIY meter designs require a housing and can be bulky when you include a battery compartment. I wanted to place the battery and trim pot along with the sensors in a compartment in my scrubber and just pour potting resin all over the LCD screen and board (using a mould of course).

Is there any reason not to do this? I understand some resins can be reasonably exothermic when curing but that can be avoided depending on which formulation you use.

Thanks,

Rob

teomannaskali

thats what im planning on doing

but, i'll ave the electronics and battery isn a separate compartment.

or even better you can have them potted but inside the canister aswell.

you may want to look at

the rebreahter Juha built.

www.breathewise.n3.net

Keep us informed on what you do

Teoman

Scuba_Vixen

If you're talking about just PO2 metering, like the Mark Munroe (PPO2.com) units, it just isn't practicle.

Those particular Datel DPMs can't reference the analog signal to supply voltage negative, so you'd need 2 power (battery) wires, cell wires depending on number of cells, and then there's the trim pot wire issue! .... Another bunch of wires, all of which are low level and wanting to pick up noise and cause error.

You'd be better to build a op-amp circuit that converts the cell millivolts to a higher level, using a 2V range DPM and then 1.000 PO2 would be 1.000V. You could use the 5V supply DPMs, (same power supply for the op amp and display) as you can reference the analog voltage to power negative. (saves a conductor in the cable) You could calibrate the DPM once and then seal it. The op amp board would reside with the battery and the cal pots would finely adjust the gain of the op amp circuit.

The only other way would be more complex and require a microcontroller so you could use software calibration. That would still require "buttons", but they can be magnetically seperate from the potted components.

The HSE PPO2 meter (uses the same case as the HS Explorer) is that type of implimentation readily available.


Good luck,


Darlene

whoizrob

I appreciate thte input but I will say that the knowledge base here concerning electronics is a bit above me. If I understand you correctly the low signal voltage of the O2 sensors is easily interferred with when running multiple long lays of wire parallel with your signal wire. I see how placing an Op Amp into the ciruict and raising the voltage would help alleviate some of those problems but that's about the limit of what I understood.

I'd love to learn more concerning circuit design (that is, after all, why I started this rebreather project, to learn as much as I could along the way!). Do you have a suggestion of some good reference material I can look into regarding Op Amps and circuit design? I seem to recall my old Ham radio license required some circuit building knowledge but that was Eons ago =)

Thanks again!

Rob

Scuba_Vixen

Most all the component manufacturers have about a bazillion app notes and technical publications regarding the types of components they make.

TI, Fairchild, National, ST Micro, etc. have tons of stuff on their websites, whether you want to build op amp designs or ADC stuff, you can get details you wouldn't imagine.

Darlene

whoizrob

I've been digging around and found quite a bit of info on op amps. Fun stuff and I must say I am just getting the hang of it all and will need some quality time with a breadboard to get it right. Just curious why you suggest an op amp versa a standard low voltage amplifier? Also, what op amp do you prefer, the 714 or? The local Radio shack also stocks the National tl82s.

If I understood you correctly then (using an 8 pin 741 op amp) I'd send power (pins 4/7) and cell signal (pins 2/3???) to the op amp. I'd solder a trim pot across the the op amp ( pins 1/5) to fine tune the output signal (pin 6). I'd then run the output signal along with battery negative to the DPM?

Thanks for all the help. I love this stuff =)

Rob

Scuba_Vixen

The 741 or any of the other "jelly bean" op amps are not well suited for precision and low signal circuits.

You want what's called a single-supply, rail to rail, precision op amp.... which means the amp output can swing fully from 0V to the supply voltage. (typically 5V) It also responds very linearly from very low input signals to full supply voltage input signals. (as long as signal swing and gain are correct)

Typical examples in the 8 and 14 pin thru hole packages would be OPA2340 or MAX 494.

There's a lot more to this than I can explain in a short post, but what I'm talking about is a non-inverting op amp amplifier circuit from a single supply, (5V) with the - cell referenced to ground (the neg supply connection). The way you fine tune the output is by using a pot as about 25% of the resistance value for the total from the inverting input to ground. Effectively you're adjusting the gain of the amplifier.

If you use 2V display modules instead of 200mv like the PPO2.com meters, you can set the amp gain to ~ 20 (and then fine tune it from there), and the meters would read .210 for an input to the amp of 10.5mv

The idea of using the amp is so you don't have to put pots on the display module board to back door adjust the gain of the display by changing the reference voltage.


I just finished the amp board for my project, so I'll post a pic shortly.


Darlene


Adit: Here's the amp board ready to be potted. ...... And with the resin just poured.

Genesis

Yep - the LM324 style Opamps are ok if you're going to do software on a PIC, as they are low drift, but have an annoying habit of having an offset which varies from unit to unit (within the same package too!) so you have to null that out in your code.

Their linearity is quite good, but the offset is murder if you're trying to do a meter-style thing as opposed to a microprocessor-type.

For that sort of thing Darlene's got the right idea... and the precision units aren't bad in any event - just a bit more money.

__________________
"A venturesome minority will always be eager to get off on their own, and no obstacles should be placed in their path; let them take risks for Godsake, let them get lost, sunburnt, stranded, drowned, eaten by bears, buried alive under avalanches - that is the right and privilege of any free American."
http://www.denninger.net
http://www.diversunion.org/liability.htm - Fix the Diving Cert racket

rdmmdr

the last lm 324 i used had an offset of 0 on one channel 24mv on the next and 36mv on the third. I spent an hour looking for stray noise. genesis has the circuits on his site. there are several pics out there that come with the support board included that you could use. i have used the oopic and the lcdx for projects. the last project is using the lcdx using two of its analog channels for the switches and the extra digitals for the i2c comm to the head pic. that way i have only four wires from the head to the handset. if you want to incorporate data logging or deco you will need to upgrade the eproms to 64k. am trying to get it working right now think i have a problem with the code, on the the second pic. need to put it on the scope and see who is talking to whom . the lcdx comes with a screen, and eight analog channels built in for about 90.00 it is a little big but can be shortened with different standoffs.

the basic designs are as follows

1. handset reads the values directly from the sensor, hand set comprised of mv meters.
pros simple easy and cheap
cons no hud, possible noise from unamplified design, no solenoid control

2. small pic controller mounted behind the lcd display at the hand set. reading the sensors only.
pros simple easy and cheap
cons no hud, no redundancy and still have to build the op amp for the analog channels.

3. small pic controller mounted behind the lcd display at the hand set. reading the sensors and having outputs for hud and or solenoid
pros cheap
cons no redundancy and still have to build the op amp for the analog channels. lots of wires between the head and the handset.

4. two pics with displays, heads and independent batteries, with hud and soleniod and all the extra bells and whistles you wish to program
pros redundacy, full eccr control, can be set for data logging and or deco
cons extra cost, no longer simple to program, extra hardware involved.

my design is based around number 4.
hand set
lcdx controller with 20x4 display
pressure sensor
two mag reed switches
the reed switchs and magnets come frome the alarm store, the springs come from the bolt shop along with stainless clevis pins for the buttons and stainless 7/16 nf plug drilled through for the retainers.

head unit consists of three boards
the op amp, the pic for the a/d and outputs, and a power board for the 5 volts to the controllers and 6 volts for the solenoid. stacked to fit in a 1.5 inch tube.

the hud is a three color three wire led red/green/amber connect to a head

the solenoid connection is the same as the hud on the other head
the heads and handsets are interchangeable. just swap the software

the batteries are 9.6v nmhi from rc cars,cut apart and made round, cheap and easy to come by.

everthing is potted so no cases to leak.

the only plug is from the battieries to the head, use crous hinds here they are larger then the fishers but they are rated to 10,000 psi ,pre molded, and cheaper. the plug acts as the on off switch. the wps or the ws series is what you want.

data logging to the eprom will limit the life of the unit by the number of times you write to it so be carefull when you program not to get in a loop that will burn the eprom. accessing your data should be done through the rs232 port this will allow you to bring the connections to the surface of your cases and not worry about sealing them, four wire, tx/rx/atn/gnd. there are several good programs that will load to an excel sheet for downloads.

make sure that you use an absolute pressure guage because the gauge type are referenced to the inside of your cases.

for the deco use what ever model you are comfortable with but check it hard against the model.

date hold batteries are nice but the unit will run for several days on the large batteries and it only has to be set once. plug the batteries in on friday and they will run for at least five days before you need to recharge.

cables
four wire twisted pair shielded water proof comm cable is available from ammrom by the foot, they also have the battery connectors. if it works for the sat rats it will work for this.

vr3

you can also include a vr3 port so you dont need the extra sensor by just using the extra op amp of a quad op amp set to 0 gain. (See i got it right this time darlene.) and a fisher connector.

case material

read darlene's saga of the leaking cases.

alarm conditions

single sensor failure,dual sensor failure, complete sensor failure, current limiting, and fialure of the solenoid all need to be addressed. my software puts the unit into manual mod at any of these failuresactivates the hud alarm the handset buzzer and requires a dil flush. that should allow you to determine the cause of the problem. and take action. current limiting is the tough one to find but pressurizing the head to 1.6 atm durring o2 cal will solve this one. so design your heads to take the pressure. sensor polling has been discussed here so read up.

i am sorry for get long winded, hope this helps with your design ideas.

rick m